Dissertation - Michael Becker

Turkish, some of them being very familiar words, and some being obsolete words thatwere not familiar to the speakers we consulted.The materials were recorded in a sound attenuated booth into a Macintosh computer at a44.1 KHz sampling rate. Using Praat (Boersma & Weenink 2008), the token judged best ofeach suffixed form was spliced and normalized for peak intensity and pitch. Peak intensitywas normalized using Praat’s “scale peak” function set to 0.6. For pitch normalization,three points were manually labeled in each affixed form: the onset of the word, the onsetof the root’s final segment (the onset of the burst in the case of stops), and the offset of theword. Then, a reversed V-shaped pitch contour was superimposed on the materials, with apitch of 110 Hz at the onset of the word, 170 Hz at the onset of the root-final segment, and70 Hz at the offset of the word. These values were chosen in order to best fit most of thespeaker’s actual productions, such that changes would be minimal.Finally, for each stimulus, two .wav files were created by concatenating the two suffixedforms with a 0.8-second silence between the two, once with the voiceless form followedby the voiced form, and once with the voiced followed by the voiceless. A linguist whois a native speaker of Turkish verified that the final materials were of satisfactory quality.While she had some concerns about stress being perceived non-finally in a few of the filleritems, no problems were found with the stimuli.2.3.1.3 ProcedureBefore the beginning of the experiment, speakers were reminded that voicing alternationsare lexically-specific by presenting a familiar non-alternating paradigm (top ∼ top-u‘ball’) next to a familiar alternating paradigm (Ãep ∼ Ãeb-i ‘pocket’). Then, speakers wereasked to choose the possessive form of two familiar alternating nouns (dolap ‘cupboard’and aaÙ ‘tree’), and feedback was given on their choices.The stimuli were presented in a self-paced forced-choice task. The base form, e.g. fetwas presented in Turkish orthography, which reflects the relevant aspects of the phonology35